Re: Duplicate Item Pointers in Gin index

On Wed, Feb 21, 2018 at 12:31 PM, Peter Geoghegan <pg(at)bowt(dot)ie> wrote:
> On Tue, Feb 20, 2018 at 7:19 PM, Masahiko Sawada <sawada(dot)mshk(at)gmail(dot)com> wrote:
>> IIUC, ginInsertCleanup() holds ExclusiveLock on metapage during adding
>> tuples in the pending list to the accumulator. And inserting entries
>> to the pending list also requires ExclusiveLock on metapage. This
>> behavior is not relevant with that bug fix. So I don't think that
>> backend2 can inserts a tuple while backend1 is processing the pending
>> list.
>
> You mean that you think that the problem that Siva described cannot
> happen once Postgres has commit
> 3b2787e1f8f1eeeb6bd9565288ab210262705b56?

It's true the problem cannot happen once postgres has that commit.
Since the commit 3b2787e fixed so as not to happen the #1 event that
Siva described, so the problem cannot happen once postgres has it. But
what I meant is that I think the event #3 and #4 cannot happen even
without that commit.

Regards,

--
Masahiko Sawada
NIPPON TELEGRAPH AND TELEPHONE CORPORATION
NTT Open Source Software Center